Big Data Lessons from the 2012 Election

by | Feb 14, 2013 | Video Programming Insights

Last month, Barack Obama was inaugurated for a second term as president. This success was owed partly to a complete re-imagining of how modern political campaign should be run, specifically considering the use of technology. Earlier this month, as reported in the previous article, media companies tossed around ideas as to how to shift to digital distribution. While television tries to find its future, a hint of the path forward can be gleaned by examining how Obama for America changed its culture and thought process to revolutionize elections.

The flurry of post-election reports revealed that the Obama and Romney campaigns had two distinct approaches to technology and the understanding of its power. Romney for President saw technology as the means to an end, a tool to be used. Obama for America dramatically altered its campaign to fit the emerging tech world. In the end, much of Obama’s electoral and fundraising success was tracked back to the software that glued the campaign together. Coincidentally, Romney for President’s strategic election technology—entitled Orca—crashed on election day.

This is a narrative best viewed as a lesson on meticulous planning versus destructive assumption. And while impulses of schadenfreude may steer us towards examining the destruction, success is a better thing to understand and mimic.

How to Embrace an Unknown

Ten days after the election, The Atlantic published “When the Nerds Go Marching In” by Alexis C. Madrigal, which examines the personalities behind Obama for America and its outlook on technology. From the beginning, the strategy was to fully integrate new data processes and social media into the campaign. Narwhal, the name of the platform, was a large network that included every bit of data and resource for the campaign, including “what Obama for America knew about voters, canvassers, event-goers, and phone-bankers.”

The campaign hired Harper Reed as their CTO, and a former Microsoft product manager to manage Narwhal. From there, Narwhal was treated as a tech product. According to John Slaby: “The real innovation in 2012 is that we had world-class technologists inside a campaign.” A lot has been written about how Obama for America used data when it came to mining and targeting voters, but this mindset change is equally as important. While the goal was still to win a presidency, the 2012 iteration treated data as something to be fully integrated, not just given a passing glance.

Madrigal offers a comparison: Similar to the 2008 Obama campaign, Mitt Romney’s campaign treated technology as a means to an end. It was seen as a way to assist and advise the campaign, not something that was an integral part of the campaign. In 2012, Reed instituted simulations of a Narwhal network collapse, and his team troubleshooted how to keep the system up and running. On Election Day 2008, the Obama for America election app Houdini crashed just like Orca did. But this time around, Obama’s Narwhal stayed up and running.

Benefits of Learning the Unknown

With this change of mindset, Obama for America took optimal advantage of data collection. The Wall Street Journal reported that the campaign dropped web cookies on as many computers as possible, using them to gather information for strategic email marketing and website development down the line. The 2012 Obama for America analytics department—his time lead by Chief scientist Rayid Ghani, formerly of Accenture—was quadruple the 2008 department.

Web tracking took a back seat to election and Democratic party registration. ProPublica’s reported on the extensive amount of data collected. Obama for America compiled and compared magazine and cable subscribers lists to registered voter lists to determine which ad placements would be the most cost-effective. The lists were also matched with television viewing patterns from Rentrack and FourthWall Media, giving Obama for America insight into the bests times for reaching certain voter demographics. Voters were categorized beyond party affiliation and political beliefs—by the probability of being influenced to vote for the Democratic ticket. This “persuadability” was then matched with viewing information to give a complete picture as to which times and channels would have the biggest net vote benefit for the campaign.

A new type of campaigning emerged from the combination of these tactics. By aggregating as much data as could be found, Obama for America knew exactly whom to target, when to target them, and the nature of their biggest electoral concerns.

The Takeaway

The big data tactic that was used in the 2012 election can be replicated in almost any consumer field. (Ghani had previously worked with grocers to research what qualities of orange juice were more appealing to shoppers.) The biggest hurdle for media companies is to compile the resources and technology to effectively implement this strategy—because moving into an online realm requires a thought process completely different than that which has worked for the past seven decades.

In The Atlantic article, Alexis C. Madrigal compares campaigns to short-lived start-up companies, where little job security and a quick turnaround means little time to build on previous years. Obama for America treated the 2012 campaign as a completely reinvigorated business to be built from the ground up, in pursuit of familiar goals through a new approach.

With innovation needed in media distribution, it’s important that companies look outward and invest in technology and the right personnel. A forward-thinking solution has a greater chance of being created by the top talents in technology. Over the next few years, the most successful media companies will be those that create intelligent solutions within new frameworks, not outdated ones.

By Zac Stockton, Product Manager


Share This